[sharp-discuss] [vonrhein] Re: mir in autosharp

[Clemens Vonrhein]

Dear Sameeta,

>        I am having problms in doing MIR with
> autosharp. As was written in the manual that  there
> maybe problems of handedness and origin shifts if all
> the derivatives are taken atonce, i tried SIR with
> each of the three derivative sets that i had i.e. Hg
> Th and Ur. I got one one site for Th and Hg with the
> CC (test) of 0.12 and 0.08 respectively. Refinement of
> the heavy atom positions followed by density
> modification gave the CC of the maps to be 57.33%(for
> the Th site )and a CC of 80.33% for the Hg site.

Single site SIR(AS) is always a bit tricky: 

 - which spacegroup do you have?

 - do you have anomalous data?

 - did you specify to look only for a single site?

 - what does the automatic interpretation of the residual maps say?


What I would do is to have a look at the residual maps for each of
your Th and Hg SIR(AS) run (go to the last SHARP log file in your
initial/original hand): do these residual maps look flat or are there
additional sites? I would be suprised, if you had only a single site
for all your derivatives. Try to add additional sites (with peak
heights > 6 sigma). You can do this from the residual map plotting
control panel (through the peak picking button).

But be careful in certain spacegroups: because it's SIR(AS) with a
single site you might have a mirror plane going through your first
site. In this case add just a single additional site first and have
another look at the residual maps to add all minor sites.

Of course, all this you will have to do outside of autoSHARP: use the
'Start SHARP' button on the main SHARP Control Panel (based on
new_whatever.i).

>       Now when i try to use two derivatives together
> and putting the site for only one of them I get the
> following problems:

So you're using 

  - autoSHARP with MIR (2 derivatives)

  - one of the derivatives has a whatever.hatom file

  - for the second you want to search for sites


> 1) It rejects the site that was input after the
> weeding step(it does that for both the sites though
> both have good enough initial CC(test) values and give
> maps with good enough CC's. Why does it do like that.
> 2) The program then goes ahead and finds a site for
> the other derivative with the help of the program
> RANTAN. This is the same site it found when SIR was
> done.
> now when it comes to sharp refinement, the pkmaps logs
> file shows an error saying "unable to find operator
> with MAXLVL or COMBine card" what is the meaning of
> this error message and what has to be done .

Ok - shouldn't do that. The problem is that I never tried MIR in
autoSHARP and giving some (but not all) derivatives a known set of
sites (via *.hatom files). It works if giving known sites for all
derivatives (but then you need to be sure that everything is on the
same origin). 

What I would do is the following:

 - refine a SIR(AS) with the best derivative (or try all of them) -
   see above!

 - use a MTZ file which contains all your derivative data

 - after solvent flattening you can calculate ordinary difference
   fourier maps to see possible sites in all derivatives (present in
   MTZ file): these are on same origin/handedness

The main point is to do SIR(AS) untill you're happy with the residual
maps (no minor sites) and you have the best possible SIR(AS) phases.

Hope that helps.

Cheers

Clemens

-- 

***************************************************************
* Clemens Vonrhein, Ph.D.          vonrhein@GlobalPhasing.com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park     Tel: +44-(0)1223-353033
*  Cambridge CB3 0AX, UK           Fax: +44-(0)1223-366889
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************